mloss.org new software

MLPACK 1.0.6

Wed, 19 Jun 2013 19:21:07 -0000

MLPACK is a scalable C++ machine learning library. Its aim is to make large-scale machine learning possible for novice users by means of a simple, consistent API, while simultaneously exploiting C++ language features to provide maximum performance and maximum flexibility for expert users.

The following methods are provided:

Density Estimation Trees
Euclidean Minimum Spanning Trees
Fast Exact Max-Kernel Search (FastMKS)
Gaussian Mixture Models (GMMs)
Hidden Markov Models (HMMs)
Kernel Principal Components Analysis (KPCA)
K-Means Clustering
Least-Angle Regression (LARS/LASSO)
Local Coordinate Coding
Locality-Sensitive Hashing (LSH)
Naive Bayes Classifier
Neighborhood Components Analysis (NCA)
Nonnegative Matrix Factorization (NMF)
Principal Components Analysis (PCA)
RADICAL (ICA)
Rank-Approximate Nearest Neighbor (RANN)
Simple Least-Squares Linear Regression
Sparse Coding
Tree-based Neighbor Search (all-k-nearest-neighbors, all-k-furthest-neighbors), using either kd-trees or cover trees
Tree-based Range Search

Command-line executables are provided for each of these, and the C++ classes which define the methods are highly flexible, extensible, and modular. More information (including documentation, tutorials, and bug reports) is available at http://www.mlpack.org/.

SonS and MDSonS, Software for hierarchical clustering visualization V1

Jose Maria Martinez Martinez — Tue, 18 Jun 2013 12:18:05 -0000

This toolbox consist of two Matlab functions "sons.m" and "mdsons.m". To learn how to use these functions (and to interpret these methods), check the folder named "Tutorial" and click on the "HELP" html file. Two cases of study are detailed in this file. Also you can replicate the results and content of html help file. For this purpose, there are two folders named "SonS(Synthetic data set)" and "MDSon_German_Elections" in which the two cases of study proposed in the html help file are presented. To do this, you must run the programs in the following order:

1) clustering.m: The program carries out a hierarchical clustering. It saves the centroids and the labels of the data of the different hierarchies (the labels indicate the belonging to each cluster).

2) example.m: This program calls the sons.m and mdsons.m functions for visualizing the results of the clustering provided by "clustering.m". "sons.m" and "mdsons.m" functions represent only one hierarchical level, so in "example.m" there are different calls of these functions in order to represent the different hierarchies.

Darwin 1.6

Stephen Gould — Tue, 18 Jun 2013 07:39:41 -0000

The framework includes a wide range of standard machine learning and graphical models algorithms as well as reference implementations for common applications particularly in computer vision. It also includes Matlab wrappers for many components of the library and an experimental GUI for designing machine learning data flows.

GPstuff 4.2

Mon, 17 Jun 2013 13:22:52 -0000

GPstuff is a toolbox for Bayesian Modeling with Gaussian Processes with following features and more:

Several covariance functions (e.g. squared exponential, exponential, Matérn, periodic and a compactly supported piece wise polynomial function)
- Sums, products and scaling of covariance functions
- Euclidean and delta distance
Several mean functions with marginalized parameters
Several likelihood/observation models
- Continuous observations: Gaussian, Gaussian scale mixture (MCMC only), Student's-t, quantile regression
- Classification: Logit, Probit, multinomial logit (softmax), multinomial probit
- Count data: Binomial, Poisson, (Zero truncated) Negative-Binomial, Hurdle model, Zero-inflated Negative-Binomial, Multinomial
- Survival: Cox-PH, Weibull, log-Gaussian, log-logistic
- Point process: Log-Gaussian Cox process
- Density estimation and regression: logistic GP
- Other: derivative observations (for sexp covariance function only)
Hierarchical priors for hyperparameters
Sparse models
- Sparse matrix routines for compactly supported covariance functions
- Fully and partially independent conditional (FIC, PIC)
- Compactly supported plus FIC (CS+FIC)
- Variational sparse (VAR), Deterministic training conditional (DTC), Subset of regressors (SOR) (Gaussian/EP only)
- PASS-GP
Latent inference
- Exact (Gaussian only)
- Laplace, Expectation propagation (EP), Parallel EP, Robust-EP
- marginal posterior corrections (cm2 and fact)
- Scaled Metropolis, Hamiltonian Monte Carlo (HMC), Scaled HMC, Elliptical slice sampling
Hyperparameter inference
- Type II ML/MAP
- Leave-one-out cross-validation (LOO-CV), Laplace/EP LOO-CV
- Metropolis, HMC, No-U-Turn-Sampler (NUTS), Slice Sampling (SLS), Surrogate SLS, Shrinking-rank SLS, Covariance-matching SLS
- Grid, CCD, Importance sampling
Model assessment
- LOO-CV, Laplace/EP LOO-CV, IS-LOO-CV, k-fold-CV
- WAIC, DIC
- Average predictive comparison

If you use GPstuff, please use the reference: Jarno Vanhatalo, Jaakko Riihimäki, Jouni Hartikainen, Pasi Jylänki, Ville Tolvanen, Aki Vehtari (2013). GPstuff: Bayesian Modeling with Gaussian Processes. In Journal of Machine Learning Research, 14:1175-1179.

See also user guide at http://arxiv.org/abs/1206.5754

ClusterEval 1.0

chris de vries — Sun, 16 Jun 2013 04:15:30 -0000

This program compares a clustering to a ground truth set of categories according to multiple different measures. It also includes a novel approach called 'Divergence from a Random Baseline' that augments existing measures to correct for ineffective clusterings. It has been used in the evaluation of clustering at the INEX XML Mining track at INEX in 2009 and 2010, and the Social Event Detection task at MediaEval in 2013.

OpenOpt 0.50

Dmitrey Kroshko — Sat, 15 Jun 2013 17:09:30 -0000

Universal Python-written numerical optimization toolbox. Problems: NLP, LP, QP, SDP, SOCP, DFP(Non-linear Data Fit), NSP(nonsmooth), MILP, LSP, LLSP, MMP, GLP, MINLP, MOP etc. Connects to dozens of solvers (some are C- or Fortran-written).

Provides graphic output of convergence and some more numerical optimization "MUST HAVE" features.

Our another tool FuncDesigner allows to involve automatic differentiation, uncertainty and interval analysis, categorical variables, general logical constraints, more convenient modeling of some optimization problems, systems of (non)linear equations (solver "interalg" can find ALL roots), possibly sparse/overdetermined, systems of ordinary differential equations and much more.

AISAIC 1.0.0610

Guoqiang Yu, Bai Zhang, Xiguo Yuan, Xuchu Hou — Thu, 13 Jun 2013 21:54:55 -0000

AISAIC software for analyzing human DNA copy numbers and detecting significant copy number alterations

MultiBoost 1.2.01

Wed, 12 Jun 2013 22:34:04 -0000

AdaBoost [Freund-Schapire, 1997] is one of the best off-the-shelf supervised classification methods developed in the last fifteen years. Despite (or perhaps due to?) its simplicity and versatility, it is suprisingly under-represented in the family of open softwares. The goal of this submission is to fill this gap.

Our implementation is based on the AdaBoost.MH algorithm [Schapire-Singer, 1999]. It is an intrinsically multi-class classification method (unlike SVM for example), and it was easy to extend to multi-label or multi-task classification (when one item can belong to several classes). The program package can be divided into four modules that can be changed more-or-less independently depending on the application.

The strong learner. It tells you how to boost. The main boosting engine is AdaBoost.MH, but we have also implemented FilterBoost for a research project, and Arc-GV which is basically straightforward once you have the main engine. Other possible strong learners could be LogitBoost and ADTrees.
The base (or weak) learner. It tells you what features to boost. Right now we have two basic (feature-wise) base learners: decision stumps for real-valued features and indicators for nominal features. We have two meta base learners: trees and products. They can use any base learner and construct a generic complex base learner using a "classic" tree-structure (decision trees), or using the product of simple base learners (self advertisement: boosting products of stumps is the best reported no-domain-knowledge algorithm on MNIST after Hinton and Salakhutdinov's deep belief nets). We have also implemented Haar filters [Viola-Jones, 2004] for image classification, a meta base learner that uses stumps over a high dimensional feature space computed "on the fly". It is a nice example of a domain dependent base learner that works hand-in-hand with its appropriate data structure.
The data representation. The basic data structure is a matrix of observations with a vector of labels. We also have multi-label classification when the label data is also a full matrix. In addition, we have sparse data representation for both the observation matrix and the label matrix. In general, base learners are implemented to work with their own data representation (for example, sparse stumps work on sparse observation matrices, or Haar filters work on a integral image data representation).
Data parser. We can read in data in arff and svmlight formats.

The base learner/data structure combinations cover a large spectrum of possible applications, but the main advantage of the package is that it is easy (for the advanced user) to adapt MultiBoost to a specific (non-standard) application by implementing the base learner and data structure interfaces that work together.

The source code is available from the website multiboost.org. It can be compiled on Mac OS X, Linux, and Microsoft Windows. The interface is command line execution with switches.

Information Theoretical Estimators 0.39

Zoltan Szabo — Wed, 12 Jun 2013 13:12:52 -0000

ITE can estimate

entropy: Shannon entropy, Rényi entropy, Tsallis entropy (Havrda and Charvát entropy), complex entropy,
mutual information: generalized variance, kernel canonical correlation analysis, kernel generalized variance, Hilbert-Schmidt independence criterion, Shannon mutual information, L2 mutual information, Rényi mutual information, Tsallis mutual information, copula-based kernel dependency, multivariate version of Hoeffding's Phi, complex mutual information, Cauchy-Schwartz quadratic mutual information, Euclidean distance based quadratic mutual information, distance covariance, distance correlation, approximate correntropy independence measure,
divergence: Kullback-Leibler divergence (relative entropy, I directed divergence), L2 divergence, Rényi divergence, Tsallis divergence, Hellinger distance, Bhattacharyya distance, maximum mean discrepancy (kernel distance, an integral probability metric), J-distance (symmetrised Kullback-Leibler divergence, J divergence), Cauchy-Schwartz divergence, Euclidean distance based divergence, energy distance (specially the Cramer-Von Mises distance), Jensen-Shannon divergence, Jensen-Rényi divergence, K divergence, L divergence, certain f-divergences (Csiszár-Morimoto divergence, Ali-Silvey distance), non-symmetric Bregman distance (Bregman divergence), Jensen-Tsallis divergence, symmetric Bregman distance,
association measures, including measures of concordance: multivariate extensions of Spearman's rho (Spearman's rank correlation coefficient, grade correlation coefficient), correntropy, centered correntropy, correntropy coefficient, correntropy induced metric, centered correntropy induced metric, multivariate extension of Blomqvist's beta (medial correlation coefficient), multivariate conditional version of Spearman's rho, lower/upper tail dependence via conditional Spearman's rho,
cross quantities: cross-entropy.

ITE offers solution methods for

Independent Subspace Analysis (ISA) and
its extensions to different linear-, controlled-, post nonlinear-, complex valued-, partially observed models, as well as to systems with nonparametric source dynamics.

ITE is

written in Matlab/Octave,
multi-platform (tested extensively on Windows and Linux),
free and open source (released under the GNU GPLv3(>=) license).

APCluster 1.3.2

Ulrich Bodenhofer, Andreas Kothmeier, Johannes Palme — Wed, 12 Jun 2013 11:38:01 -0000

Affinity propagation (AP) is a clustering algorithm that has been introduced by Brendan J. Frey and Delbert Dueck. The authors themselves describe affinity propagation as follows:

"An algorithm that identifies exemplars among data points and forms clusters of data points around these exemplars. It operates by simultaneously considering all data point as potential exemplars and exchanging messages between data points until a good set of exemplars and clusters emerges."

AP has been applied in various fields recently, among which bioinformatics is becoming increasingly important. Frey and Dueck have made their algorithm available as Matlab code. Matlab, however, is relatively uncommon in bioinformatics. Instead, the statistical computing platform R has become a widely accepted standard in this field. In order to leverage affinity propagation for bioinformatics applications, we have implemented affinity propagation as an R package. Note, however, that the given package is in no way restricted to bioinformatics applications. It is as generally applicable as Frey’s and Dueck’s original Matlab code.

The package further implements leveraged affinity propagation, exemplar-based agglomerative clustering, and various tools for visual analysis of clustering results.